function y=multvtmcmc(mu,choltr,v) 
y=mu+(choltr*randn(length(mu),1))*sqrt((v/chi2rnd(v)));
% ========================================================================
% function y=multvtmcmc(mu,choltr,v); 
% Draw a sample y from a multivariate t-Student 
% t(mu,sigma,v)  v is degrees of freedom > 2 
% mu=mean vector 
% sigma=choltr*(choltr') 
%       choltr=transpose( chol(sigma ) ); 
% v=degress of freedom 
% Following the notation in Gelman et al. Page 481 
% Mean(y)= mu         
% Var(y) =(v/(v-2))*sigma
%
% To obtain choltr=C in mcmc sampling use VAR2TCHOL.m 
% See TEST_T.m for a debug
% V must be > 2 
% Alejandro Justiniano May 1 2007         
% =========================================================================

